Chapter 2 Wartime Trade Theory

security externalities of trade

Security externalities are the additional benefits arising from a state’s use of gains from trade to increase its military capabilities.⁵ International trade produces security externalities by two established pathways. First, trade provides products and foreign currency, which can be used to augment a state’s war-fighting potential. For example, the import of arms or ammunition immediately increases a state’s military capabilities. The import of food, coffee, cloth for uniforms, medical supplies, or rubber, while not directly increasing military capabilities, can affect the war effort. The export of goods furnishes a state with foreign currency, which can be used to purchase military supplies. The second pathway through which international trade produces security externalities is specialization. International trade allows states to focus on producing the items they are most efficient at making, which increases the income derived from and the productive efficiency of the domestic economy.⁶ The additional income can be used to invest in military capabilities. The freed-up resources, no longer used to manufacture products imported from abroad, can be channeled toward military production or even greater revenue generation for the purchase of military supplies.⁷

For a state to benefit from security externalities, the gains from trade have to reach the government, as it is the government that allocates funds for state security. If the government, itself or through state-owned enterprises, purchases or sells specific items abroad, the products and foreign currency that cross national boundaries accrue to the state directly. This is a straightforward case of how gains from trade can be used to augment the military capabilities of a state. However, most of the time, domestic firms, not the government itself, import and export products. To invest in military capabilities, the government needs to repurchase products from domestic merchants or expropriate them. Likewise, the government needs to tax international trade and domestic production to benefit from the gains from trade. For a state that does not tax its domestic economic activity or its trade, security externalities hardly matter.⁸

Considering how gains from trade reach the government reveals a third, previously unrecognized pathway for the generation of security externalities. When the military forces face a shortage, the state can prevent necessary products from leaving the country. That is, the state can force domestic firms to forgo the economic profits of selling at higher world prices and instead to resupply the military. Doing so allows the state to pay less for the (more efficiently generated) products and ensures that the needs of the military are prioritized over the economic benefit of merchants.

Traditionally, security externalities of trade are measured with income gains made by states engaged in trade compared to autarkic production.⁹ This measure considers the security externality of the entire trading relationship. Thus, it misses the variation between products in the severity of the generated security externality.

Focusing only on income gains from trade means ignoring the speed at which security externalities accrue. Importing a finished product of immediate benefit on the battlefield creates an immediate security externality. On the other hand, importing a raw material that takes several months to manufacture into a finished product that will be useful on the battlefield only leads to security externalities in several months. This is the basis of the first theoretical insight that helps solve the fundamental problem of trade with the enemy: security externalities are not instantaneous. Products vary temporally in terms of when security externalities become problematic.

Additionally, products vary in the amount of revenue they contribute to a state’s GDP. When more steps of the production supply chain take place domestically, the state can tax the manufacturing process more and thus receive more revenue from the circulation of the product in the domestic economy. The import of a raw material to be processed domestically contributes more to the GDP than the import of a finished product. This is the basis of the second theoretical insight that helps solve the fundamental problem of trade with the enemy: security externalities are not equally high for all products. Products vary in the amount of military capabilities that can be derived from their continued trade.

first theoretical insight: conversion time

Security externalities are not instantaneous. It takes time for a state to convert the gains from trade, whether from imports or exports, into military capabilities usable on the battlefield. More importantly, this conversion time varies by product. The shortest practical conversion time is selling a gun to the enemy across battle lines. Though incredibly rare, such situations have occurred throughout history. At the beginning of the siege of Grave in the Wars of Louis XIV, the French commander of the town, at the instruction of the secretary of state for war, sold gunpowder to the Dutch, who were besieging them.¹⁰

A particularly long conversion time would be associated with the import of raw materials that are processed domestically for export. For example, a state imports raw gemstones and converts them into fine jewelry, which is sold to customers abroad. The conversion time would include time to make the deal for purchasing and delivering the gemstones, time to transport the gemstones to the factory, time to refine and set them into the final jewelry piece, time to transport the jewelry to the buyer, and time to receive the payment for the sale. At the end of this process, the state would still have to convert the revenue from taxing this domestic manufacturing process into military capabilities, necessitating additional time to purchase arms or food rations and deliver them to the battlefield. While even trade in raw gemstones can be used to aid the war effort, converting them into military capabilities takes a lengthy period of time.

A product’s conversion time into military capabilities is determined by four processes: transaction, transportation, production, and taxation. The length of time each process takes is both historically contingent, as each process has become faster over the centuries of warfare, and contingent on the efficiency of the state. However, regardless of the time period and the specific state, different products take varying amounts of time to circulate through the economy, leading to variation in conversion times.

Transaction time refers to the time it takes to place an order with an existing supplier in a foreign state or the time to negotiate a new contract with a different supplier. Where the government does not purchase these products itself from foreign sources, the time it takes to repurchase (or confiscate) products from domestic merchants also increases the transaction time. Time to redirect resources from civilian into military production would also fall into this category.

Transportation time accounts for all the time it takes for products to reach their destinations, including the initial import time, the shuffling of products through domestic factories, and the ultimate delivery to the battlefield. Transportation time is shortest when a state imports finished military capabilities, like arms and ammunition. In such cases, this process is only the length of time it takes to move products to the battlefield. Of course, geography plays a large role in extending or shortening this process. In a hypothetical war between Russia and Japan, selling Russia a gun at Vladivostok has a much lower transportation time than selling Russia that gun in St. Petersburg. A product’s transportation time is extended if it requires domestic manufacturing. For example, a raw material can be imported into a state, transported to a factory for manufacturing, and shipped to the battlefield. The transportation time would be the sum of these three segments. A product’s conversion time is extended each time it has to be transported to a new point for further processing.

Production time covers the time it takes to manufacture a specific product. Within the same supply chain, the more a product is processed domestically, the longer its production time. Raw cotton’s production time into a military uniform is necessarily longer than a bolt of cloth’s because the cotton’s production time has to include the processing of the raw material into a bolt of cloth, which is then processed into the uniform. The production time for exported products includes the domestic manufacture time of converting raw materials into the products being exported. The production time for imported products includes the domestic manufacture time of the imported good into the product that is consumed domestically, exported abroad, or shipped to the battlefield.

Finally, taxation time refers to the time it takes to extract taxes from the circulation of the traded products through the domestic economy. This process depends on the efficiency of the state at generating revenue through taxation. States can tax trade through import and export taxes. They can also tax the circulation of products in the domestic economy through income taxes, corporate taxes, and sales taxes, among other things. The time it takes the state to extract revenue from taxing trade increases the length of time it takes to convert the gains from trade into military capabilities. On the other hand, if the state controls aspects of trade directly, by purchasing resources or running state-owned enterprises, taxation time does not affect those products’ conversion times.

One more aspect contributes to the estimation of conversion time: strategic stockpiles. Stockpiles push back the starting time of when security externalities start accumulating. If a state has a month’s supply of gunpowder for a company of troops, any additional gunpowder imported for a month will not have direct security externalities for the current war as the troops are already well outfitted. A state can still convert the gunpowder into military capabilities by treating it as a tradable commodity; however, with the presence of stockpiles, the conversion time increases. Since the gunpowder is not useful for its intended military purpose—as there is already a sufficient supply—the state can resell the gunpowder to a third party and use the proceeds to purchase food rations for the troops, increasing its conversion time by the transaction, transportation, production, and taxation times associated with the transfer. Only after the strategic stockpiles have been used does the conversion time of gunpowder to military capabilities revert back to the time it takes to import and transport it to the battlefield.

When making wartime commercial decisions, a state does not attempt to measure, down to the second, the amount of time the opponent requires to purchase a gun, process sulfur into gunpowder, or transport cannons to the battlefield. It would take a herculean effort and more time than it is worth to determine the conversion time with that degree of precision. States do, however, have general intelligence about the length of manufacturing processes and the presence or lack of transportation infrastructure in places where goods can be bought and sold. They make educated guesses, from information collected by embassies, about the levels of stockpiles in enemy states and the alternative trade deals firms in those states could make. While they cannot assign products explicit conversion times, states can create a ranking of products based on how long their conversion into military capabilities is likely to take. Beyond that, state decisions are colored by the accuracy of their information. During the Crimean War, British decision-makers, expecting the Russian aristocracy to amass all the profits, assumed that the Russian government received almost no revenue from trade.¹¹ Because of this misunderstanding, Britain could assess trade with Russia during the war to be fairly safe from security externalities. Wartime commercial decisions can be, and usually are, amended when states realize that their initial information is faulty.

Different products traded between two enemies can be ranked on the basis of their approximate conversion times into military capabilities. A product closer to the end of a supply chain has a faster conversion time than a product at the beginning of the same supply chain. Crude oil will always have a longer conversion time than diesel because diesel is of immediate utility on the battlefield and crude oil has to be processed into a usable fuel type. A finished good that in itself increases military capability (guns, ammunition, tanks, fighter jet fuel, etc.) has a faster conversion time than finished goods that are not immediately useful on the battlefield. However, it is not necessarily the case that products at the beginning of supply chains in military production have faster conversion times than products at the beginning of supply chains in nonmilitary production. This comparison depends on how long it takes to get through the entire supply chain. It takes less than a day to process wheat into bread, and it takes about twenty-two months to go from a metal bolt to an F-35.¹² Even if the process of sowing grain and reaping the wheat is considered, it still takes longer to manufacture an F-35 than to make a loaf of bread.

second theoretical insight: loss of revenue

Just as products vary in the amount of time it takes for their trade to generate security externalities, they also vary in the severity of security externalities created. A product’s relative importance to the economy is associated with the contribution the product’s circulation in the domestic economy makes to the state’s GDP. This ultimately affects the investment the state can make into its military capabilities. Thus, losing trade in a product that generates considerable revenue for a state would carry a greater security externality than losing trade in a product that generates next to no revenue. Traded products, whether imports or exports, can be ranked on the basis of the revenue lost if a state severs trade.

Severing trade in products with close substitutes or sources of supply other than the enemy, regardless of how important these products are to the economy of the state, would not have a large impact on the continuity of funds available.¹³ If the product offered by the alternate source is identical to the one being traded with the enemy, then changing trading partners only requires a one-time switching cost. There might be additional costs if the product is of a different quality or comes at a different price; however, compared to severing trade in the product outright, switching to a similar product or another source only slightly affects the revenue of the state.¹⁴ For example, while the export of hydrocarbons, like coal or oil, tends to contribute greatly to a state’s GDP, the near ubiquitous global demand for them, as well as the increased domestic wartime demand, generally ensures the availability of alternative customers and the continuation of the stream of revenue even after trade with the enemy is prohibited. The impact of severing trade with one country on the GDP generated by a state’s hydrocarbon industry is likely to be minimal. Severing trade with the enemy in products with close substitutes, likewise, would not greatly affect the revenue of the state or the state’s ability to invest in its long-term security. As long as the new imported product allows the industry to continue functioning, the contribution to the GDP should be minimally affected.

If a product is neither substitutable nor available from alternative sources of trade, a prohibition on trade will translate into a loss of the full contribution to GDP of the product’s circulation in the domestic economy. The amount of revenue lost can be ranked on the basis of the number of points at which the specific product is taxed—import tax, corporation tax if the product is involved in domestic production, export tax, personal tax if it is imported and sold by individuals, sales tax if it is sold domestically after import, and so on. While the relative rates of various taxes influence which products contribute more to the revenue extracted by the state, in general, the more points at which a product is taxed, the more revenue it brings to the state. The import of a motorcycle, a finished good, contributes less to GDP than the import of steel, a raw material used in the domestic car-manufacturing industry. Whereas the motorcycle can only be taxed at the moment of importation and resale to a consumer, the steel can also be taxed through each step of the manufacturing process until it is finally sold as a finished car. Similarly, the export of raw materials contributes less to the GDP of the state than does the export of finished goods of the same supply chain. If the export of a domestically manufactured product is prohibited and there is no demand for the product domestically, in the long term, its production will cease completely. This will deprive the state of the tax revenue from the manufacturing industry. In general, products that require domestic manufacturing contribute more to GDP than products exported to be manufactured elsewhere.

One more aspect contributes to the severity of a product’s security externality: whether the product is involved in the arms industry. Even when weapons manufacturing is a small portion of the overall economy, states treat this industry as more important than its contribution to GDP would suggest. Losing trade in products in the arms industry has a double effect: decreasing the revenue of the state and preventing the state from manufacturing its own military capabilities. As a result, the import of products related to the arms industry is weighed more heavily than just the revenue lost from severed trade. The export of products related to military capabilities, on the other hand, does not receive this additional scrutiny. In war, there tends to be an increased domestic demand for military capabilities, which can substitute for export markets.

Fortunately, states tend to be keenly aware of where their revenue comes from. The state knows which of its industries are most important to the domestic economy and has access to domestic experts to ascertain which products are required to keep those industries running. Using their trade records with the enemy and with the rest of the world, states can determine which products are substitutable and which can be gained or sold elsewhere. Thus, the state can rank products on the basis of the expected revenue loss from severed trade, from products the state can afford to lose to those that will only be sacrificed to the cause of immediate survival.

The likely arrangement of products suggested here, on the scale measuring the loss of state revenue due to severed trade, is based on an efficient economy. However, even if the economy is not operating efficiently, the state can create a similar scale, though products would be located differently on it. For example, if prohibiting trade in a product increases domestic production to a level that increases a state’s GDP—that is, the war permits the state to escape the unfavorable economic policies that have been imposed on it—severing trade will have a positive impact on state security, and trade in this product will be on the lowest end of the scale. If the state does not tax trade, then only the product’s circulation in the domestic economy affects the state’s revenue. Alternatively, if the state does not tax domestic production, then only the import or export taxes are relevant for determining where a product will rank. Regardless of how specifically its economy functions, the state can still sort products on the basis of the revenue lost from severing trade and therefore their impact on security.

solution to the fundamental problem of trade with the enemy

These two theoretical insights provide a solution to the fundamental problem of trade with the enemy. Trade in all products does not have to be severed for the state to obey its security imperative. Likewise, trade in all products does not need to continue during the war for the state to follow its economic imperative. The goal of the security imperative is to protect the state from increasing the enemy’s military capabilities during the war. The state prohibits trade in products that benefit the enemy in the current war and allows trade in products the enemy will not have time to convert into military capabilities. Trade in products with short conversion times will likely be severed; trade in products with long conversion times will likely be allowed. The goal of the economic imperative is to ensure the state continues the stream of investment into military capabilities. The state can afford to prohibit trade in products that do not contribute greatly to the economy but allows trade in products whose loss would lead to substantial harm to the GDP. Trade in products that contribute greatly to the domestic economy will likely be allowed.

There are two potential areas of overlap between these two rationales for allowing trade with the enemy to continue during the war. First, products could have long conversion times while their loss would not do much or any damage to the state’s GDP—that is, the security imperative would not require severing trade, but the economic imperative would not require protecting the trade either. This is easy to reconcile. So long as there is no military reason to sever trade, the state does not need to give up even a small amount of revenue. Trade in these products can continue as long as the enemy does not have time to convert them into military capabilities. Second, products could have short conversion times while their loss would be catastrophic for the domestic economy—that is, the security imperative would dictate that trade be severed, but the economic imperative would mandate that it must be maintained. Here the choice is between aiding the current war effort and ensuring the state can continue to defend itself in the future, assuming it survives the current war. Products that fall into this category can only be derived from the enemy and are key inputs into important industries on which the economy of the state rests. The state will most likely err on the side of allowing such trade to continue to avoid Pyrrhic victories; the state cannot afford the revenue loss from severed trade if it is to maintain its long-term security.

Figure 2.2 is a visual representation of the solution to the fundamental problem of trade with the enemy. Each dot denotes a product traded between a state and its enemy before the war started. The y axis represents the revenue loss from severing trade in a specific product. The state can easily afford to lose trade in products at the bottom of the axis. Products at the top are too important to the domestic economy for the state to sever trade in them. The x axis represents the enemy’s conversion time of a specific product into military capabilities. Products at the left side of the axis are too dangerous to trade with the enemy during the war. Products on the right side are unlikely to aid the enemy in the war.

For example, for Russian considerations of trade with Japan during the Russo-Japanese War (1904–5), guns would be located in the lower left corner of the chart. Japan could convert guns to military capabilities very quickly, and the export of guns was not a very important industry for the Russian economy. Moreover, there would be an increased domestic demand for guns to make up for the loss of foreign sales. Steel would be located farther to the right and higher than guns. It would take longer to process steel into military capabilities, and Russia was a steel-exporting country.¹⁵

The curves symbolize some general patterns for a state’s wartime commercial policy. Products within the inner curve are most likely to be prohibited from trade with the enemy. The enemy can quickly use them to affect the outcome on the battlefield, and the state making the wartime commercial policy can easily afford to lose the trade. Products close to the outer curve are least likely to be prohibited from trade. Either the enemy will not have time to militarily benefit from the trade or the state cannot afford to lose the trade despite the enemy’s military benefits.

doesn’t a state care about an enemy’s revenue from trade?

When assessing the security externalities of trade, the state prioritizes the enemy’s conversion time. The state focuses on winning the current war, which requires preventing the enemy from increasing its military strength on the battlefield. Any gains the enemy might acquire for its long-term security are discounted; while a state cares deeply about its own long-term security, the enemy’s is considerably less relevant. Because of this, the state is not as concerned about the amount of revenue the enemy receives—that is, the importance of the traded products to the enemy’s economy—as it is about the speed with which the enemy receives it. Providing the enemy with a billion dollars or with raw materials that allow it to generate a billion dollars, if it will have access to the sum only after the war is over, is not that problematic.¹⁶ After all, it hardly matters that the enemy could have added infinite resources to its war effort if it never has the chance to do so.

Granted, these resources will be available to the enemy after the war ends, but calculations about gains from trade are different in peacetime than in wartime. First, the winner of the war can impose terms on the defeated party that include reparation payments or otherwise prevent the enemy from using the gains from trade to increase their military capabilities. Second, the enemy’s next conflict does not have to be with the same state. The enemy could fight a third party and use any gains from trade made during the previous war against this third party in the next one.¹⁷ Third, in peacetime states have to pay greater attention to relative gains vis-à-vis third parties. Trade between the state and its enemy benefits both sides. If, however, the enemy receives the same product from a third party, it is the enemy and the third party that benefit. The state that severed trade with the enemy would be left at a relative loss compared to the enemy and the third party. Maintaining trade with the enemy during the war maintains trade ties and prevents relative losses compared to neutral states; this continued trade becomes even more beneficial in peacetime.

doesn’t a state care about its own conversion times?

When the state calculates its own security externalities from trade, it prioritizes the revenue lost from severing trade over conversion time. The focus is on guaranteeing the continuity of investment into military capabilities over the long term. Because of the long-term outlook, the state can discount the speed with which revenue is generated to ensure that the revenue never stops. Losing out on a billion dollars today does not matter as much if the state can still receive the billion tomorrow. The immediate effect of revenue lost from severing trade can be ameliorated by drawing funds from alternate sources. If the revenue is made up at a later date, the flow of investment into the security of the state need not be affected. However, if the revenue loss is permanent, the long-term security of the state is compromised.

If a state needs to increase its military capabilities for the current war, it can turn to several sources of revenue: for example, income taxes or loans from foreign governments.¹⁸ The absence of one source of revenue—the gains from trade—can be dealt with in the short term. During a war, a state has greater discretion in diverting funds toward the war effort. Although it would be advantageous to trade with the enemy to receive gains from trade that can be converted into military capabilities for the current war, it is not necessary. Alternative sources of funding can be used to pay for short-term increases in military capabilities to gain the battlefield victory. Because a state cares deeply about its long-term security, it can accept some delays to ensure there are no permanent interruptions in the investment into its future security.

expected length of war

The expected length of war is the decision-makers’ estimate about how long the current war will last: Is the state expecting a quick victory or facing a prolonged struggle? This estimation can be based on a number of considerations. One such factor is the balance of capabilities between the major belligerents. A state with an overwhelming military advantage, even if only assumed, will likely expect a short war. A state believing itself to possess weaker capabilities than the enemy will likely expect a lengthy war.²⁰ Another factor is the development of new technologies or strategies predicted to provide an edge in fighting.²¹ For example, the Germans in World War II developed a rapid blitzkrieg strategy that increased their confidence in a quick victory.²² A third possibility is the use of analogy or knowledge gained from previous conflicts. If a state is facing a war that seems to have similar conditions to a war in recent memory, the past war could influence the expectations of the length of the coming one.

Empirically, states have proved notoriously bad at assessing the expected length of war. There is a tendency to overestimate the probability of a quick decisive blow to the enemy and a tendency to underestimate the need to think through the consequences of subsequent actions if the initial knockout blow fails.²³ Falling prey to the allure of battle, states assume that they can win a quick victory, even though most of the wars fought in human history have been won by attrition.²⁴ Thus, it is not uncommon for states to start their wars with the expectation of a short struggle to be ended with a decisive battlefield victory.²⁵ No one wants to self-select into a long war. The one notable exception is perhaps colonial wars, where colonial forces expected long, drawn-out fights given their acknowledged lack of capabilities.

While states typically tend toward underestimation, they may also overestimate the length of the war. France prepared for a prolonged battle of attrition against Germany at the start of World War II and was unpleasantly surprised by a quick defeat. The United States and its coalition allies expected a quick victory in the Gulf War of 1991 and won an even quicker victory after only one hundred hours of fighting.

A state’s assessment of the length of war can, and frequently does, change during the war. States with overly confident assessments of the length of war have to update their expectations when the war lasts longer than the initial estimate. Since states have incentives to misrepresent their relative strength,²⁶ a state overestimating the enemy’s strength would plan for a longer war than is necessary. As the war reveals private information, states update their expectations about the length of the war.²⁷ The optimism for the success of a fresh offensive can decrease the estimate of the time left until the end of the conflict, while the failure of an offensive or a fresh counteroffensive by the enemy can increase the assessment.

The belligerents involved in the war need not agree on the expected length of war. It is possible that one side plans for a prolonged conflict while the other expects a quick victory, as was the case between France and Germany in World War II. It is, likewise, possible for both sides to expect a quick victory, as was the case for the same belligerents in World War I. Each state makes its own calculations about the expected length of the coming war.

What is considered by belligerents to be a short or a long war is historically contingent. Before the invention of railroads, when an army had to march on foot to the battlefield, a short war would have been much longer than a short war in the age of cargo trucks and airplanes. For World War I, which was expected to be short by most belligerents, short meant that the “boys would be home by Christmas,” after around five months.²⁸ Several modern wars, on the other hand, are measured in days, not months. For these wars, an expectation of a five month war might seem quite long. In the cases examined in the following chapters,²⁹ a short war refers to a conflict that lasts no more than one campaign season and does not require troops to winter near the battlefield. A long war lasts more than one campaign season and does require troops to winter near the battlefield.

There are two sources of information that can help establish the state’s expectation of the length of war. The first comes from the public statements of leaders and their private correspondences with advisers and military commanders. What sort of war were the leaders preparing to fight? How long a campaign were they expecting? How optimistic were they about achieving their goals in the set time frame? How much logistical support did they allocate? Did they plan to call up reservists? The second source of information is the actual arrangements made for the war effort: Did the state make costly investments into the war, suggesting they were preparing for a long conflict? For example, the construction of extensive fortifications, the preparations of trenches before the war begins, or the construction of field railroads to solidify supply lines would indicate a state preparing for a long war. All of these actions would be a waste of resources and effort if the state expected to win the war before such construction was even complete. On the other hand, sending troops to a location with a cold climate without winter clothes indicates preparation for a short war.

expected stakes of war

Expected stakes of war refers to the decision-makers’ assessment of how existentially threatening the current war will become. This helps the state navigate the trade-off between dedicating all of its efforts to winning the current war at the cost of undermining its future security and dedicating all of its efforts to long-term security at the expense of potentially losing the current war. Future security refers to the state’s ability to defend itself after the current war. However, if the current war becomes particularly prolonged and engrossing, then future security can also encompass the ability of the state to prosecute the expected future years of the war at the same time as defending itself from other threats. A permanent interruption in the stream of revenue invested in the state’s military capabilities is detrimental to the long-term security of the state. Severing trade in products essential to the economy thus places the state’s future security at risk. Concurrently, if this trade helps the enemy increase its military capabilities in the current war, the trade makes the current war harder to win. Thus, the state has to make an intertemporal trade-off. The value states attach to the stakes of the war determines the choice between future security and current survival. If the war is expected to be existentially threatening, the state will prioritize current survival and will be willing to bear greater revenue loss from severed trade. On the other hand, if the war is not over vital national security concerns, the state can prioritize long-term security and will be willing to bear less revenue loss. The more at stake in the war, the more states will be willing to sever trade in products necessary for key domestic industries, decreasing investment in their future security, to prevent their opponents from benefiting militarily from trade, thus increasing their own chances of winning the current war.

As with the expected length of war, the expected stakes can change during the conflict. Factors marking the war’s increasing stakes could include the fighting shifting closer to the state’s capital, increasing fears of decisive defeat on the battlefield; the opponent adopting maximalist war aims; or the opponent exhibiting ideological fanaticism. All of these factors would make the state more willing to sacrifice some aspect of its future security to increase the chances of winning the current war. On the other hand, when additional allies join the war effort, each member of the expanding alliance might feel a lessening of the intensity of war, making them less willing to sacrifice various aspects of trade.

Two sources of information can help establish the state’s expectation of the stakes of war. First is the leaders’ assessment of the reasons the war is being fought, their war aims, and the vigor with which these goals will be pursued. Is state death a possible outcome? Does the war involve an invasion of the state’s homeland or a fight in a distant region? Is the war fought over a vital national security interest or a matter of less strategic importance?³⁰ The closer the struggle comes to affecting the survival of the state, the greater the expected stakes of the war. The second source of information comes from the state’s preparations for war on the home front. If the state is preparing large-scale evacuations of citizens or devising a rationing scheme, the expected stakes of war are likely to be high. Likewise, if the state is preparing to mobilize the entire economy for the war effort, it is expecting an intense war. On the other hand, if the state commits only a small portion of its full military capabilities, it is likely expecting the stakes to be lower. If the majority of the citizens are unaware that their state is fighting a war, it is likely not existentially threatening.

how restrictive should the wartime commercial policy be?

The state’s expectations about the coming war determine how much trade the wartime commercial policy suppresses. In a short, peripheral war, the state is likely to develop a lax wartime commercial policy, exemplified by the top portion of figure 2.3. In a short war, the enemy will only have time to convert a fairly low number of products into military capabilities; therefore, only a few products need to be prohibited from trade with the enemy. The expectation of lower stakes in the war suggests the state will prioritize long-term security, avoiding large revenue losses and thus continuing trade with the enemy in many products.

In a long, existential struggle, a state is likely to have a restrictive wartime commercial policy, exemplified by the bottom portion of figure 2.3. A lengthy war provides the enemy with considerable opportunities to convert gains from trade into military capabilities, so a state will sever trade in a considerable number of products to avoid giving the enemy an advantage. High stakes of war motivate states to discount the future and sacrifice considerable revenue to make the current war marginally easier to fight. Even trade in products essential to the state’s economy could be prohibited from trade to ensure the survival of the state.

War expectations change during the war. A state expecting a short victory can soon discover it is in for the long haul. A state expecting a homeland invasion can manage to repulse the enemy still on distant shores. Updated war expectations lead to changes in the wartime commercial policy. How much the policy suppresses wartime trade varies directly with the state’s expectations about the coming war.

Furthermore, the two war characteristics do not necessarily move in tandem. States can maintain expectations of a short war while increasing the expected stakes, as the French experienced in the Franco-Prussian War of 1870–71. The government of the newly minted Third French Republic refused to accept that France could lose the war, especially so quickly, until Paris fell.³¹ However, the expected stakes of the war increased considerably when the German armies laid siege to the capital city. Similarly, states can update the expected length of war without changing their expectations about the stakes, as the French did in the Algerian War of Independence of 1954–62. Even as the expected length of war increased, the expected stakes remained stable.

is it not endogenous?

Some observers might raise concerns over endogeneity, as the prohibitions on trade could plausibly affect the expected length of war. This is not a problem for this theory for two reasons.

First, a prohibition on trade with the enemy is a long-term policy instrument. The effects of the policy are felt only after it has been enforced for some substantial period of time. After the enemy feels the effects, the economic pressure increases proportionally with the length of time the policy is enforced. In the short term, however, a prohibition on trade does not have an effect. The enemy may have strategic stockpiles of prohibited products and thereby may not feel the effects of severed trade. Likewise, states have alternative sources of income, so a lack of profit from prohibited trade can, in the short term, be made up with alternative methods of war funding. It is only once the initial stockpiles of products (or profits) are exhausted and the supply of additional trade is severely reduced that the economic pain of trade sanctions can be felt.³⁴ Thus, an initial expectation of a short war cannot stem from the chosen wartime commercial policy.

Second, if the expectation of the length of the war and the prohibitions on trade influence each other, no decisions on wartime trade could ever be reached. Consider one traded product that is on the threshold of a war length expectation. If there were an endogeneity problem, prohibiting trade in this product would shorten the expected length of war. However, lowering the expected length of the war also means trade in the product in question should be allowed. If the product is traded with the enemy, the war length expectation rises again, and so the product should not be traded with the enemy. This cycle continues indefinitely. The very fact that decisions on wartime trade are made shows that the judgments about the length of war and the prohibitions on products are made separately.

A second endogeneity concern might be raised over the effect of peacetime activity on wartime commercial policies. For instance, firms can respond to geopolitical tensions by redirecting their trade away from the expected enemy to safer countries, reducing the amount of trade between the two future belligerents while still in peacetime.³⁵ This theory explains state preferences for wartime trade, not individual firm behavior or the direct bilateral trade flows between states. But even that aside, this concern is not a problem for the theory.

First, the trade-follows-the-flag argument—firms, anticipating that war will lead to commercial disruptions, take measures to diversify their trade away from the potential enemy to maintain stable business operations³⁶—relies on the assumption that wartime trade does not exist. Both the theory and evidence presented in this book cast doubt on this assumption. States have strategic reasons to permit wartime trade in specific wars. Awareness of this pattern of behavior would have the opposite effect on firm decision-making. Wartime trade tends to be more lucrative, as the demand for products is frequently inflated. Anticipation of wartime trade could solidify trade flows with the enemy.³⁷

Second, if prewar dynamics were to remove trade with the adversary, we should expect several effects in wartime that are not borne out by the empirical evidence. If firms, independently, diversify their trade from the potential adversary, we should see few government regulations on wartime trade, as they would be redundant. In crisis situations, governments have little time or available labor to waste, much less the time and effort it takes to generate a nuanced product-level policy that is updated several times a month. Since governments produce such detailed wartime commercial policies, it suggests that firms are not independently acting to limit their own trade with the enemy.

Conversely, if firms diversify only trade in products with low adjustment costs, we should expect wartime trade with the enemy to remain mostly in products where the enemy is a predominant supplier or market. As such trade defines economic dependence, losing it would generate great costs to the state; thus, if firms conform to this behavior, we would expect considerable wartime trade across all wars. This is not the observed empirical pattern. Additionally, we would expect the majority of wartime trade, under these conditions, to be for reasons of state necessity, as opposed to the temporal mechanism proposed in the theory. This also does not conform to the observed empirical pattern.